Do Type Ia Supernovae prove Lambda > 0 ?

Boud Roukema boud_roukema w camk.edu.pl
Sob, 5 Sty 2002, 00:36:36 CET


Cze¶æ wszytkie,

On Fri, 4 Jan 2002, Andrzej Marecki wrote:

> Paper: astro-ph/0201034
http://de.arXiv.org/abs/astro-ph/0201034
> From: M.Rowan-Robinson <m.rrobinson w ic.ac.uk>
> Date: Thu, 3 Jan 2002 16:29:08 GMT   (64kb)
> 
> Title: Do Type Ia Supernovae prove Lambda > 0 ?
> Authors: Michael Rowan-Robinson
> Comments: 10 pages, 10 figures. Accepted for publication by MNRAS

Well, the principle of checking things like extinction and reanalysing
the data is a good one, especially given that just a fraction of a 
magnitude makes the difference between an Omega_m=1 and a Lambda 
model, as far as the SNeIa data go.

But assuming that all RR's corrections regarding extinction are
correct, the real result of the paper (section 7.(4)) is just that the
rejection of the Omega_m=1 model is "only" 2.8-4.6\sigma (instead of
6-10(?)\sigma).

Well,
3\sigma (double sided) is 99.7%
4\sigma (double sided) is 99.994%

Still strong evidence in favour of non-zero \Lambda!

Our constraint (RMB02, the proofs are still sitting on my desk... mea culpa)
http://de.arXiv.org/abs/astro-ph/0106135
rejecting Omega_Lambda=0 is itself "only" 3\sigma.  Our result is totally 
independent of the SNeIa results, it's a standard ruler, not a standard
candle, and there are also the weak gravitational lensing constraints,
plus the very nice result from our friend Roman Juszkiewicz, simply
from galaxies' relative velocities, constraining Omega_m=0.35\pm 0.15 
(if large-scale bias is zero)
http://cdsads.u-strasbg.fr/cgi-bin/nph-bib_query?bibcode=2000Sci...287..109J&db_key=AST&high=3bd83e24b015894
so while RR is right that people should remain skeptical and to suggest
that a sociological warning is a good idea, his analysis does nothing
to suggest that the present "fashion" is wrong. 

And I think the referee should have hassled him on his Hubble diagram.
(Fig 10)

* He said that he himself finds z=1.4 for the SN 1997ff, but he plots
it for z=1.7. If he trusted his own analysis, he would have put the
spot at z=1.4, much closer to the (0.3,0.7) curve.

* He didn't put *error bars* on the high z points of Fig 10. The error
bars put by Riess et al. 2001 on SN 1997ff are *big*, particularly in the
magnitude direction. The uncertainties in modelling gravitational lens
time delays are still big - and should have been plotted.

* It's rather unphysical to write "v" for velocities greater than
c. "zc" is correct, redshifts can be equal to or greater than 1. But
velocities of massive objects cannot be greater than or equal to c,
unless you redefine velocity as v= d_(d_comoving)/d_(t_cosmological),
but this is a definition which would only have very special uses in 
restricted contexts, and certainly not on this diagram.

* He should have labelled "luminosity distance", not just "distance".

So you can relax :-), Andrzej: (0.3,0.7) is still the best local
cosmological parameter estimate to use when looking for global
cosmological parameters!

Pozdrawiam,
Boud

 


Więcej informacji o liście Cosmo-torun